Re: [PATCH: gdb/mi + doco] -var-update

[Nick Roberts <nickrob at snap dot net dot nz>]

Daniel Jacobowitz writes:
 > On Sat, Jul 16, 2005 at 03:12:00AM +1200, Nick Roberts wrote:
 > >  > BTW, if it's OK with you, I would prefer that you add new testcases
 > >  > rather than modifying existing ones.  Yes, the ones that are there are
 > >  > slightly redundant.  But changing what a test case is testing is bad
 > >  > form for long-term results analysis.
 > > 
 > > OK.  I'm not very comfortable with expect, but I'll give it a go.
 > 
 > I have learned a frightening (to me anyway) amount about expect and TCL
 > lately; bug me with any questions you may have.

Thanks.

 > >  > Now that this patch is in, I believe your current gdb-mi.el will work
 > >  > with CVS GDB.  Is that right?  If so, could you post it?  I promise to
 > >  > look at it promptly.
 > > 
 > > Yes it does work, but not perfectly (if it did it could go straight into
 > > the Emacs CVS repository).  It has changed quite a lot since last time.
 > > Unfortunately it still requires Emacs in CVS, but a release is planned.
 > 
 > Confused - I thought we'd decided it belonged with GDB, not with
 > emacs, because of the GDB version dependencies?

The eventual aim is to move it to Emacs CVS.  While MI and gdb-mi.el are
changing , GDB CVS seems a good place for it.  You're right though, since
it currently requires MI features that are only in GDB CVS, it wouldn't
make sense to move it to Emacs CVS yet.

 > I won't try to second-guess you about this code - my LISP sucks.  But
 > could you explain briefly why you changed from -break-insert to "break"
 > in some places?  Yet at the same time changed from until to
 > -exec-until...

I changed "until" to -exec-until as it's an MI command.  I changed
-break-insert back to "break" because the output from -break-insert spills
into the GUD buffer.  Clearly the proper solution is to handle the output
of -break-insert, but I haven't got that far yet.  I don't know how
important that is because gdb-mi.el will always have to be able to handle
CLI user input from the GUD buffer anyway.

 > Anyway, go ahead and check this in.

Committed.  Thanks.

Nick